144
Acc. Chem. Res. 1996, 29, 144-154
Strategy and Tactics in Combinatorial Organic Synthesis. Applications to Drug Discovery E. M. GORDON,* M. A. GALLOP,
AND
D. V. PATEL
Affymax Research Institute, 4001 Miranda Avenue, Palo Alto, California 94304 Received September 26, 1995
I. Introduction The single greatest time constraint on the process of drug discovery has been the practice of synthesizing and evaluating candidate molecules one at a time. One vivid illustration of this approach was the global effort expended over a 30 year period in which thousands of β-lactam compounds were serially synthesized and screened for antibiotic activity.1,2 During the 1970s, a more detailed appreciation of enzyme mechanism heralded the advent of rational strategies for drug design,3-5 culminating in the creation of the antihypertensive drug captopril as one notable example.6,7 The utilization of structural information derived from crystallographic and NMR studies of target biomolecules, together with modeling and computational methods, has further increased the sophistication of the rational design approach, and collectively these techniques have somewhat reduced the number of compounds needed to be made in order to find and optimize a drug lead. Although dramatic technological advances in modern biology have uncovered many new biochemical pathways and targets for pharmaceutical intervention, the rate of discovery of new drugs has not increased commensurately. Recent revolutionary changes in the health-care sector have compounded the problem by increasing pressure to shorten drug discovery and development time lines. The pharmaceutical industry has reacted by urgently seeking new schemes for increasing productivity and for more quickly capitalEric M. Gordon was born in New York City in 1946. He was graduated with a B.S. from Beloit College (1967) and received M.S. (1969) and Ph.D. (1973) degrees from the University of WisconsinsMadison. Dr. Gordon was a postdoctoral fellow at Yale University (1973−1974) and at The Squibb Institute for Medical Research (1974−1975) in Princeton, NJ. He spent 17 years at Squibb and Bristol Myers Squibb, most recently as Director of Medicinal Chemistry. In 1992 Dr. Gordon joined the Affymax Research Institute as Vice-President of Research and Director of Chemistry. His research interests include drug discovery, the rational design of enzyme inhibitors, the chemistry of amino acids, peptides, and natural products, and combinatorial chemistry. Dr. Gordon has authored 170 papers and patents in these areas. Mark Gallop was born in Whangarei, New Zealand, in 1962. He received B.Sc. and M.Sc. degrees in chemistry from the University of Auckland. He obtained his Ph.D. degree in inorganic chemistry from the University of Cambridge in 1988, working with Prof. Jack Lewis. He was a Lindemann postdoctoral fellow investigating antibody catalysis in the laboratories of Profs. Peter Schultz and Robert Bergman at the University of California, Berkeley. He joined the Affymax Research Institute in 1990, where his research has focused on the development of combinatorial chemical technologies. Since 1995 he has been Director of Combinatorial Chemistry at Affymax. Dinesh V. Patel was born in Baroda, India, in 1957. He received his B.Sc. (1976) in industrial chemistry and M.Sc. (1978) in organic chemistry from S. P. University in India. He obtained his Ph.D. (1984) in organic chemistry from Rutgers University under the guidance of Prof. Spencer Knapp. After completing postdoctoral research (1984−1985) with Prof. Charles Sih at the University of WisconsinsMadison, he joined the Medicinal Chemistry Department at The Squibb Institute for Medical Research (1985) in Princeton, NJ. He spent 8 years at Squibb and Bristol Myers Squibb and became a Group Leader in Medicinal Chemistry. He joined Affymax Research Institute in 1993, where he is currently Director of Combinatorial Drug Discovery Chemistry. His research interests include the rational design of enzyme inhibitors, the chemistry of amino acids, peptides, and natural products, and applications of combinatorial chemistry to drug discovery.
0001-4842/96/0129-0144$12.00/0
izing on the steady stream of potential new drug targets (fueled, in part, by the fruits of the human genome project). Combinatorial chemistry (for historical background, see ref 8) has emerged as a long-term solution to the high costs of serial organic synthesis of drug leads. Around 1990, the idea was crystallizing in the minds of a number of researchers that if sound synthetic strategies could be devised, the necessary chemical tools might already exist to efficiently generate large collections of molecules for biological evaluation.10-19 The newly-formed chemistry group at Affymax became excited by the power inherent in creating many small, nonpolymeric organic molecules simultaneously.8,9 Strategies were defined so that this wealth of molecular diversity not only would contain the familiar pharmacophores of the past but also could lead to the pharmacophores of the future.20 II. Strategy in Combinatorial Synthesis Comparison of the conventional approaches to drug discovery with emerging combinatorial methods reveals an apparent discontinuity in strategies and tactics which must be brought to bear. Though the principles underlying chemical reactions are of course invariant, the practice of combinatorial organic chemistry diverges markedly from serial compound synthesis. Whereas synthesis in medicinal chemistry conventionally serves the goal (usually) of producing a single product of previously specified structure for bioassay, the goal of combinatorial chemistry is to (1) Cephalosporins and Penicillins; Chemistry and Biology; Flynn, E. H., Ed.; Academic Press: New York, 1973. (2) Beta-Lactam Antibiotics; Chemistry and Biology; Morin, R. B., Gorman, M., Eds.; Academic Press: New York, 1982; Vol. 1-3. (3) Hirschmann, R. Angew. Chem., Int. Ed. Engl. 1991, 30, 1278. (4) Patchett, A. A. J. Med. Chem. 1993, 36, 2051. (5) Wiley, R. A.; Rich, D. H. Med. Res. Rev. 1993, 13, 327. (6) Ondetti, M. A.; Rubin, B.; Cushman, D. W. Science 1977, 196, 441. (7) Petrillo, E. W.; Ondetti, M. A. Med. Res. Rev. 1982, 2, 1. (8) Gallop, M. A.; Barrett, R. W.; Dower, W. J.; Fodor, S. P. A.; Gordon, E. M. J. Med. Chem. 1994, 37, 1233. (9) Gordon, E. M.; Barrett R. W; Dower W. J.; Fodor, S. P. A.; Gallop, M. A. J. Med. Chem. 1994, 37, 1385. (10) DeWitt, S. H.; Kiely, J. S.; Stankovic, C. J.; Schroeder, M. C.; Reynolds Cody, D. M.; Pavia, M. R. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 6909. (11) Desai, M. C.; Zuckermann, R. N.; Moos, W. H. Drug Dev. Res. 1994, 33, 174. (12) Zuckermann, R. N.; Martin, E. J.; Spellmeyer, D. C.; Stauber, G. B.; Shoemaker, K. R.; Kerr, J. M.; Figliozzi, G. M.; Goff, D. A.; Siani M. A.; Simon, R. J.; Banville, S. C.; Brown, E. G.; Wang, L.; Richter, L. S.; Moos, W. H. J. Med. Chem. 1994, 37, 2678. (13) Kick, E. K.; Ellman, J. A. J. Med. Chem. 1995, 38, 1427. (14) Virgilio, A. A.; Ellman, J. A. J. Am. Chem. Soc. 1994, 116, 11580. (15) Bunin, B. A.; Ellman, J. A. J. Am. Chem. Soc. 1992, 114, 10997. (16) Plunkett, M. J.; Ellman, J. A. J. Am. Chem. Soc. 1995, 117, 3306. (17) Backes, B. J.; Ellman, J. A. J. Am. Chem. Soc. 1994, 116, 11171. (18) Bunin, B. A.; Plunkett, M. J.; Ellman, J. A. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 4708. (19) Thompson, L. A.; Ellman, J. A. Tetrahedron Lett. 1994, 35, 9333. (20) Gordon, E. M. Curr. Opin. Biotechnol. 1995, 6, 624.
© 1996 American Chemical Society
Combinatorial Organic Synthesis in Drug Discovery
Acc. Chem. Res., Vol. 29, No. 3, 1996 145
create searchable populations of molecules. Combinatorial chemistry is a type of synthetic strategy which leads to collections of differing molecules, i.e., “chemical libraries”. The process proceeds by the systematic interconnection of a set, or sets, of chemical building blocks. This approach has clear parallels in Nature, where the polymerization of small monomeric units (amino acids, monosaccharides, nucleosides, etc.) produces the various classes of key biological macromolecules. Building blocks (BBs) are small, reactive molecules which may be intercombined with each other, or with members of other BB families (e.g., amines, carboxylic acids, aldehydes, etc.) in varied ways to eventually produce a combinatorial library containing many different products. Because the acquisition of BBs is a resource-intensive activity, these collections, once formed, should be fully leveraged by their integration into many varied synthetic schemes and into the production of numerous libraries. In drug discovery, at least two major types of combinatorial libraries are apparent: (1) primary libraries for use in random screening, where a priori no specific structure or substructure is obligatory as a starting point; and (2) directed, biased, or focused libraries where the library is purposefully constructed around a known starting structure having biological activity (i.e., a lead compound), to be used in optimization of the activity or other properties of the lead molecule. Approaches based on a specific structure can often be assisted by a retrocombinatorial analysis of the parent compound, in which comparative evaluation of various retrosynthetic dissections is accompanied by an assessment of whether an adequate variety of the required building blocks are readily available, and whether chemistry in the forward direction is sufficiently reliable to drive library production. While primary and focused library approaches demand rather different strategies for library construction, both must permit the extraction of information made available by library screening, i.e., the ability to know or deduce the structure of active compounds. From the perspective of drug discovery, library members should usually conform to predetermined criteria, among which are chemical stability, conformational flexibility/inflexibility, molecular weight constraints Leup > Ilep, Alap > Valp > Glyp, a result that is in agreement with the literature finding.47 In the next step of deconvolution, the best P1 residue (Cbz-Phe) (47) (a) Bartlett, P. A.; Marlowe, C. K. Biochemistry 1987, 26, 8553. (b) Morgan, B. P.; Scholtz, J. M.; Ballinger, M. D.; Zipkin, I. D.; Bartlett, P. A. J. Am. Chem. Soc. 1991, 113, 297. (48) Karanewsky, D. S.; Badia, M. C.; Cushman, S. W.; DeForrest, J. M.; Dejneka, T.; Loots, M. J.; Perri, M. G.; Petrillo, E. W.; Powell, J. R. J. Med. Chem. 1988, 31, 204. (49) Campbell, D. A. J. Org. Chem. 1992, 57, 6331. (50) Campbell, D. A.; Bermak, J. C. J. Org. Chem. 1994, 59, 658. (51) Campbell, D. A.; Bermak, J. C. J. Am. Chem. Soc. 1994, 116, 6039. (52) Campbell, D. A.; Bermak, J. C.; Burkoth, T. S.; Patel, D. V. J. Am. Chem. Soc. 1995, 117, 5381.
Combinatorial Organic Synthesis in Drug Discovery
Acc. Chem. Res., Vol. 29, No. 3, 1996 153
Figure 6. Synthetic methodology for the preparation of peptidylphosphonate libraries.
Figure 7. Design and construction of a solid phase peptidylphosphonate library.
was held constant and a Cbz-Phep-oY(5)-Z(18)-NH-resin sublibrary was assayed, now as five pools with a distinct oY component, and with 18 members in each pool. This led to identification of the leucine analog as the optimal oY(P1′) component, a result that is once again consistent with the literature data. The final Z(P2′) deconvolution involved 18 individual CbzPhep-oLeu-Z(18)-NH-resin peptidylphosphonates. Besides identifying the most potent peptidylphosphonate reported in the literature [Z(P2′) ) Ala], a number of additional active sequences composed of basic [Z(P2′) ) arginine and histidine] and neutral H-bonding [Z(P2′) ) glutamine] residues were also uncovered (Figure 8). These results were verified by solution synthesis and in vitro thermolysin I50 determinations of these inhibitors and were found to be consistent with the solid phase depletion assay data. Conclusions and Future Prospects The preparation of small-molecule libraries at Affymax Research Institute has been accomplished through the adaptation of synthetic organic chemistry to the solid phase format, guided by a strategic analysis of the principles underlying combinatorial drug discovery, and with integration of new technology tools as appropriate. Successful implementation of a combinatorial library approach has required a highly interdisciplinary research environment involving chemistry, biology, engineering, instrumentation, optics,
informatics, and other themes. The field of smallmolecule combinatorial chemistry is but a few years old, and certainly much additional technology development will be required before these approaches are routinely and widely practiced. Looking forward, we can expect that solid phase organic chemistry will become increasingly more sophisticated, offering greater synthetic flexibility and expanding the scope of structures accessible through combinatorial synthesis. Adaptation of solid phase combinatorial chemistries to run under automated instrumental control will soon become routine, and the historically distant relationship between instrumental engineering/automation and organic synthesis will narrow with increased demands for hardware to automate more complex synthetic schemes. New screening technologies emphasizing assay minaturization will be developed to capitalize on the ready availability of screenable molecules and to cope with higher throughput. Combinatorial chemistry will be integrated with the other main tools of drug discovery, including structure and mechanism based design, molecular modeling, and computational chemistry to ultimately define a more powerful discovery paradigm. The huge numbers of molecules which will be made and screened will demand a new generation of database management tools. The most effective data analysis strategies will make use of information about both active and inactive molecules, to help construct hypothetical active sites/
154
Acc. Chem. Res., Vol. 29, No. 3, 1996
Gordon et al.
Figure 8. Screening and deconvolution of a solid phase peptidylphosphonate library against thermolysin.
binding sites of potential drug targets. Analysis of library screening results in turn will drive the next round of library synthesis. Finally and significantly, combinatorial technologies may ultimately find use in surmounting some of the typical hurdles encountered in drug development (e.g., cell penetration, metabolism, bioavailability, pharmacokinetics: combinatorial drug development). Though some chemists may see combinatorial approaches as “replacing them”, in actuality the growth of combinatorial chemistry marks a new level of empowerment for the chemist. Large numbers of molecules which in the recent past might have taken a year or a career to synthesize may now be prepared in a single experiment. The availability of such
numbers of molecules and their ability to be screened for particular criteria allows researchers to expand the scope of questions they may ask and to direct efforts to more demanding tasks. The answers to these questions will certainly form the basis for new understandings of molecular behavior. We thank Drs. Alex Zaffaroni, Gordon Ringold, Russell Howard, Jeff Jacobs, Chris Holmes, David Campbell, Mike Gordeev, Martin Murphy, Beatrice Ruhland, John Schullek, Ron Barrett, Bill Dower, and our co-workers at Affymax for their contributions to understanding this new field and R. B. Morin for his critical reading of the manuscript. AR950170U
